The proposed rules for defining what Accountable Care Organizations (ACOs) are supposed to look like, and what kinds of measures they are supposed to be accountable for, have been released for public commentary by the U.S. Department of Health and Human Services (HHS).
Of note in the proposal, there is a list of 63 measures that ACOs need to report on, in order to have access to any savings from an organized patient-centered delivery of care. These measures fall into 5 different domains: (1) patient/caregiver care experiences, (2) care coordination, (3) patient safety, (4) preventive health, and (5) at-risk populations/frail elderly health.
We previously reviewed the first domain, patient/caregiver care experiences, which are based on surveys to be carried out annually by the ACO, sent to patients who had sough care there. This domain constitutes the first 7 of the 63 measures on the table for ACOs.
The biggest bulk of measures fall into the last two categories – preventive health, and at-risk populations/frail elderly health – and constitute measures 26 through 63.
All of these measures are defined by the National Quality Foundation (NQF), and most all of them are also EHR Incentive Program (“Meaningful Use”) clinical quality measures (CQMs). They are also PQRI measures (now referred to as Physicians Quality Reporting System, or PQRS). They are a mixture of process measures, as well as a few outcomemeasures.
Compared to the relatively low bar of quality measurement seen in the EHR Incentive CQMs, the ACO measures are more aggressive – for example, measurement of diabetes care, for EHR Incentive CQMs, is supposed to simply measure blood pressure, blood glycohemoglobin (HbA1c) levels, and LDL cholesterol level; the ACO measures target specific outcomes: blood pressure needs to be less than 140/90, HbA1c needs to be less than 8%, and LDL cholesterol needs to be less than 100.
Not only is the achievement bar higher than previously stated in explicit federal policy (which, arguably is a good thing), but the breadth of measures is similarly broad.
The Preventive Health domain measures (1) influenza immunization rates, (2) pneumococcal vaccination rates, (3) mammography screening rates, (4) colorectal cancer screening rates, (5) cholesterol measurement and achievement of target levels, (6) weight screening follow up plans, (7) blood pressure measurement for all, (8) tobacco use assessment and intervention, and (9) clinical depression screening.
The At Risk Populations domain has collections of measures around the following disease states: (1) diabetes care (10 measures); (2) heart failure management (7 measures); (3) coronary artery disease management (6 measures); (4) hypertension management (2 measures); (5) chronic obstructive pulmonary disease (COPD) management (3 measures); (6) frail elderly care (3 measures).
How is an ACO supposed to report these measures to CMS? Through internal systems (EHRs and other in-group Health IT systems), an ACO might be able to measure these CQM elements, but how will they be reported?
CMS is proposing using a new tool for reporting – the Group Practice Reporting Option (CPRO) Data Collection Tool. This tool does not exist yet, but is supposed to be pattered after the already-existing PQRI group-practice reporting option. In other words, there will be a separate web interface that an ACO will need to fill out, based on outputs from their internal EHR and HIT systems.
This is for the first year. Subsequently, it is hoped that EHRs used by ACOs will be able to report these measures directly to CMS (like how CMS is hoping that EHR Incentive measure reporting will be through direct data upload, rather than attestation, in subsequent program years). The particulars of this have not been specified, however.
The other question raised is about what kinds of benchmarks of performance an ACO is to be held accountable. One might be able to state that “86% of our diabetics have HbA1c levels less than 8.0%,” but what is this to be compared against? Is this good performance, mediocre performance, or below-average performance?
There are three approaches to benchmarks that might emerge: (1) comparison against other ACOs, or the local community; (2) comparison to oneself year-over-year; (3) comparison to some fixed threshold determined by policy.
Comparison against the “local community” might be difficult, especially if there is only one ACO in a local community – and this may well be the case in many instances. If no one other than ACOs are being asked to report on this long laundry-list of measures, how can “local community” comparisons be made? Will “local community” historical standards be used, coming from historic claims data, or PQRI data reported by separate practices?
ACOs stand as a promise to improve health care delivery quality to communities, and represent a shift away from simple fee-for-volume of care and towards an outcomes-based incentive (payment) approach. There are certainly many skeptics who doubt that this will actually work. However, the general environment that expects clinical quality of care, as measured by a set of process measures as well as outcomes measures, is with us to stay. Many details have yet to be spelled out – from the measures themselves, to reporting, to comparison and benchmarking – but these matters are now open for public commentary. The final rule will emerge from this process later in the year.
EHRs will figure prominently in the new world of ACOs. Beyond the relatively few measures required for EHR Incentive (“Meaningful Use”), the ACO reporting set will need to be incorporated in EHR products, as they continue to evolve into the next year. The landscape is a rapidly evolving one!
Robert Rowley, MD
Chief Medical Officer
Practice Fusion EMR